I am in the process of creating a project examining the effects of (1) belonging and (2) prosocial affirmation interventions in a higher education STEM environment.
My C.I. are intending to study ~10 courses with an average student enrollment stat being ~60/course. Each of the courses are split into two lab groups, each roughly around 30 students/lab session.
Having a control group in both lab environments is critical, but I am struggling how to limit the likelihood of these two groups communicating about the study if in the same room?
Any tips or experience with this would be greatly appreciated!
Will the treatment be delivered while the groups are separated into labs? If so, there is far less chance that the differences would be discussed among students than if the treatment is administered when they are in the presence of each other. Each lab will assume that what they are experiencing is normal and will not especially feel the need to discuss it with others in the other lab, other than what they would do normally, "I've got a really lame/good TA; how's yours?"
Why is it necessary to do the split within each of the 10 labs, where there is indeed a potential problem with treatment "leakage" or "contamination"? why not use the alternative of conducting the treatment on both halves of 5 classes, and using the other 5 as controls?
I agree with Dr. Morgan. It seems easier to use 5 courses as experimental group and the other 5 as control group.
It depends on the nature of the intervention. If it is delivered through written material (or online) you can simply distribute the different conditions randomly students in each lab group. But if it consists oral or other observable interventions, better to use different groups as suggested by previous responses.
In doing this kind of work, we've utilized both methods. As some of the previous responses suggest, if the materials are purely written, then you can get away with giving out different stimulus materials within the same section of lab. We've had students work in groups/pairs/etc. and made it clear through verbal instructions that each group is doing the same type of problem (e.g., independent samples t test) but has a different word problem. That worked well.
We have also used an intervention that was an actual group/partner activity and the control activity was comprised of the same problems but individually completed. We split students into separate rooms for that because the of the vast difference in the nature of the activity.
I think either way could work, it just takes some thinking through and logistical coordination. Also, it will depend on whether or not you want your students to know there's a difference. In the latter example I gave, students knew there was a difference because their participation was extra credit for the course. They were told there would be different activities to complete (all worth the same amount of credit). However, they were randomly assigned to either the experimental or control groups upon arrival, so they didn't know which activity they would get to do.
I hope this helps!
Can you do a multiple baseline design? Introduce the experimental group to the intervention for a period of time with continuous measurement - then introduce it to the control group (and if possible reversing the treatment group back to control). If you can do this rotation, you always have a control group and you see effects only when the intervention is applied. E.g., 4 weeks exp group, then 4 weeks apply intervention to control group, etc.
On another note, - I'm very interested in your research, and doing similar work here in Canada. I'd love to connect....
- Badges
- Science topic
- Similar topics
- Education
- Higher Education